Definition
Proportion of AI-deferred maintenance tasks where the eventual repair cost exceeded the projected cost at the original due date.
The Measurement Gap
Condition-based deferral is one of the most appealing features of predictive maintenance: the AI says the component still has life remaining, so the planned replacement is pushed to a later date, extracting more value from the component. Legacy maintenance metrics enthusiastically track the savings. Schedule compliance reports show fewer tasks; component life metrics show longer intervals; cost reports show reduced planned maintenance spend. What no conventional metric captures is what happens when the deferral goes wrong. When a deferred bearing replacement becomes a bearing failure with shaft damage, the cost escalation from a routine job to a major rebuild is invisible in aggregate cost reporting because the rebuild is coded as a new, separate event. Backlog aging metrics track how long deferred tasks sit in the queue, but they do not link the eventual repair cost back to the original deferral decision. Work order cost variance reports compare actual to estimate on the job that was eventually performed, not to the estimate of the job that was originally planned. The causal chain between "AI said defer" and "repair cost tripled" is broken across multiple work orders and reporting periods.
Decision Supported: Is condition-based deferral saving money by extracting more component life, or losing money by allowing secondary damage that turns a routine replacement into a major rebuild?
Formula
Plain Language
Identify all maintenance tasks that were deferred based on AI or condition monitoring recommendations. For each, compare the actual repair cost when the task was eventually performed to the estimated cost at the original planned date. Count how many of those deferrals resulted in an actual cost that exceeded the original estimate by more than 25%. Divide that count by the total number of AI-deferred tasks and multiply by 100. Separately, sum the dollar difference between actual and original estimated cost across all escalated deferrals to quantify the total cost of escalation.
Notation
DDER = (Escalated_deferrals / Total_AI_deferrals) x 100
Escalation_cost = Sum(Actual_repair_cost_i - Estimated_original_cost_i) for all escalated deferrals
Variables
| Variable | Definition | Unit |
|---|---|---|
| Escalated_deferrals | Deferred tasks where actual repair cost exceeded estimated original cost by more than 25% | Count |
| Total_AI_deferrals | All tasks deferred based on AI or condition monitoring recommendation | Count |
| Actual_repair_cost_i | Total cost (parts, labor, downtime) when the deferred task was eventually performed | Currency |
| Estimated_original_cost_i | Planned cost from the job plan at the original due date, before deferral | Currency |
| Escalation_cost | Aggregate dollar impact of cost escalation across all escalated deferrals | Currency |
Proposed formula, pending validation pilots.